[Chapter 1] The Nature of Viruses

 

INTRODUCTION: A BRIEF HISTORY OF ANIMAL VIROLOGY

The history of human development has been shaped by at least three major recurring elements: (1) environmental changes; (2) human conflicts; (3) infectious diseases.

 

Infectious diseases have impacted both humans and our food supply. The origins of veterinary medicine are rooted in efforts to maintain the health of animals for food and fiber production, and animals essential for workrelated activities. Control of animal disease outbreaks was not possible until the pioneering work of the late 19th century that linked microbes to specific diseases of plants and animals (see Murphy, F.A., 2012. The Foundations of Virology, for detailed discussion).

 

Many attribute the beginning of virology with the work of Ivanofsky and Beijerinck on the transmission of tobacco mosaic virus. Both scientists were able to demonstrate transmission of the agent causing disease in tobacco plants through filters that retained bacteria. Beijerinck also noted that the filterable agent could regain its “strength” from diluted material, but only if it were put back into the tobacco plants. The concept of a replicating entity rather than a chemical or toxin had its genesis with these astute observations.

 

The era of veterinary virology had its beginning virtually at the same time as Beijerinck was characterizing tobacco mosaic virus transmission. Loeffler and Frosch applied the filtration criteria to a disease in cattle that later would be known as foot and mouth disease. Repeated passage of the filtrate into susceptible animals with the reproduction of acute disease firmly established the “contagious” nature of the filtrate and provided more evidence for a process that was inconsistent with toxic substances.

 

These early studies provided the essential operational definition of viruses as filterable agents. Chemical and physical studies revealed the structural basis of viruses nearly 40 years later. In the early 20th century, use of the filtration criteria led to the association of many acute animal diseases with what were later defined as viral infections: African horse sickness, fowl plague (high pathogenicity avian influenza), rabies, canine distemper, equine infectious anemia, rinderpest, and classical swine fever (hog cholera) (Table 1.1). In 1911, Rous discovered the first virus that could produce neoplasia (tumors), and for this discovery he was awarded a Nobel Prize. This early phase of virology was fraught with skepticism and uncertainty because of the limited tools available to study and characterize the filterable agents. Experiments using filters with varying retention parameters demonstrated the existence of filterable agents of different sizes. Some agents were inactivated with organic solvents, whereas others were resistant.

 

For equine infectious anemia, the acute and chronic forms of the disease were perplexing and an unresolved conundrum. These types of apparent inconsistencies made it difficult to establish a unifying conceptual description of the filterable agents. For research on animal diseases, early workers were restricted to using animal inoculation in order to assess the impact of a treatment on any putative diseasecausing agent.

 

The logistics could be especially daunting for studies in cattle and horses. Help in providing definition to filterable agents came from the discovery of viruses that infected bacteria. Twort in 1915 detected the existence of a filterable agent that could kill bacteria. Like its plant and animal counterparts, the strength of a dilute solution of the bacterial virus could be regained by inoculating new cultures of bacteria. Felix d’Herelle also noted the killing of bacteria by an agent that he called “bacteriophage.” He defined the plaque assay for quantitating bacteriophage, a technique to enumerate virus particles based on their ability to kill cultured cells and therefore produce holes, or plaques in the cell layer that became a keystone for defining the properties of viruses.

 

The initial studies on tobacco mosaic virus led to further understanding of “filterable agents”—namely viruses. Specifically, the high concentration of virus produced in infected tobacco plants permitted the chemical and physical characterization of the infectious material. By the early 1930s, there was evidence that the agent infecting tobacco plants was composed of protein, and that antibodies produced in rabbits could neutralize the virus. The tobacco mosaic virus was crystallized in 1935, and in 1939 was the first virus to be viewed using an electron microscope. The particulate nature of viruses was now an established fact. A further advance in animal virology was the use of embryonated eggs for culturing virus in 1931. In the same year, Shope identified influenza virus in swine; in 1933, influenza virus was isolated from human cases.

 

The identification of the strain H1N1 of influenza virus in swine might be considered the first comprehensive description of an “emerging” disease in animals—that is, a virus crossing a species barrier and maintaining itself as an agent of disease in the new species. In an attempt to move away from large-animal experimentation, and to provide model systems for human diseases such as influenza, mice and rats became important tools for studying animal viruses. These advances spawned the birth of laboratory animal medicine programs that have become an essential backbone of biomedical research. The decade 1938
48 saw major advances by Ellis, Delbruck, and Luria in the use of bacteriophage to probe the mechanism of inheritance of phenotypic traits of these bacterial viruses. Advances in understanding the properties of viruses progressed much more rapidly with bacterial viruses, because the work could be done in artificial media, without any requirement for laborious and timeconsuming propagation of viruses in either animals or plants. A key concept in virus replication, namely the latent period, was defined using one-step growth curve experiments with bacteriophage (see Chapter 2: Virus Replication).

 

This observation of the loss of infectivity for a period after the initiation of the infection directed research to define the mode of replication of viruses as totally distinct from that of all other replicating entities. Animal virus studies made a dramatic shift in emphasis with the development of reliable in vitro animal cell cultures (1948
55). As a result of intensive efforts to control poliovirus infections, single cell culture procedures were defined, cell culture media were standardized, a human cell line was developed, and growth of poliovirus in a nonneuronal cell was demonstrated. These advances all permitted the development of a plaque assay for poliovirus 35 years after the concept was defined for bacteriophage. Basic studies on animal viruses that were hindered by the necessity to work in animal systems were now possible in vitro, and the principles established for bacteriophage could be explored for animal viruses.

 

The cell culture era of animal virology had begun. The advances in virology driven by human disease control efforts were directly applicable to animal virology. Bovine viral diarrhea virus was identified as a new disease-causing agent in cattle in 1946 and by the late 1950s was considered the most economically important disease of cattle in the United States. Cell culture procedures permitted isolation of the virus and the production of a vaccine by the early 1960s. Influenza virus was detected for the first time in wild birds in 1961, which led to the identification of water fowl and shore birds as the natural reservoir of influenza A viruses. An apparent cross-species incursion of a feline parvovirus variant produced the worldwide epizootic of canine parvovirus in the late 1970s. Again, standard in vitro cell culture procedures identified the new agent and soon enabled the production of an effective vaccine.

 

The entire arterivirus family (Arteriviridae) was identified in the cell culture era of virology—specifically, equine arteritis virus (1953), lactate dehydrogenase-elevating virus (1960), simian hemorrhagic fever virus (1964), porcine reproductive and respiratory syndrome virus (1991), and most recently, wobbly possum virus (2012). The discovery of human immunodeficiency virus (HIV) in 1983 attracted global attention, but the identification of simian immunodeficiency virus (SIV) shortly thereafter may ultimately be of equal importance to the eventual control of human HIV infection. The primate system has provided the animal models for studies of pathogenesis and vaccine development. Genetic analyses established that HIV-1 and HIV-2 were closely related to the SIVs present in Old World primates, and that they were independently derived via cross-species transmission of these simian viruses. The beginnings of the molecular era of virology date to the late 1970s and early 1980s.

 

Although not specifically designed for viruses, the development of the polymerase chain reaction (PCR) in 1983 had a profound impact on virus research. Cloning of nucleic acid sequences led to the first infectious molecular clone of a virus (poliovirus) in 1981. The impact of molecular techniques on virus detection and diagnostics was demonstrated with the identification of hepatitis C virus by molecular means without isolation and in vitro propagation of the virus in cell culture. Viruses that could not be easily cultured in vitro—such as papillomaviruses, noroviruses, rotaviruses, and certain nidoviruses amongst many others—could now be characterized and routinely detected by tests that specifically detect viral nucleic acid. A remarkably impressive feat spearheaded by Taubenberger and coworkers was the molecular reconstruction of an infectious virus from RNA fragments representing the pandemic 1918 strain of H1N1 influenza A virus. Dreams of recreating extinct animals by molecular techniques may be farfetched, but these techniques can identify the early precursors of currently circulating viruses. Rapid and inexpensive nucleotide sequencing strategies are again redefining virology, and whole genomic sequencing is likely to replace less exact procedures for identifying and characterizing virus isolates and strains. Metagenomic analyses that identify all nucleic acids in biological samples as well as water and soil have identified myriads of new viruses, leaving some to estimate that viruses may contain more genetic information than all other species on earth combined.

 

In the early periods of virology, the discipline was dependent upon advances in the chemical and physical sciences. Defining the characteristics of the “filterable agents” was not possible by simply observing the impact of the agent on its host. However, as time went on, viruses became tools with which to probe the basic biochemical processes of cells, including gene transcription and translation. The bacterial viruses assisted in defining some of the basic principles of genetics through the study of mutations and the inheritance of phenotypic properties. As analytical chemical procedures were developed, it was shown that viruses contained nucleic acids, and when Watson and Crick defined the structure of DNA, viruses became key players in defining the role of nucleic acids as the database for life. Progress was so rapid in the field of virology that, by the 1980s, some believed that the future value of viruses would simply be as tools for studying cellular processes. However, the unpredictable emergence of new viruses such as HIV, hepatitis C, Nipah and Hendra, severe acute respiratory syndrome coronavirus and the related Middle East respiratory syndrome coronavirus, and highly-pathogenic H5N1 influenza, together with the reemergence of already recognized viruses, such as ebola in West Africa, or their spread into previously free areas, such as West Nile virus into North America, chikungunya virus into Indian Ocean islands, Asia, and the Americas, Zika virus in the Americas, and bluetongue and Schmallenberg viruses into Europe, clearly confirm that much has yet to be learned about this class of infectious agents and the diseases that they cause.

 

Veterinary virology began as a discipline focusing on the effects of viral infections on animals of agricultural significance. Control of these infections relied on advances in understanding the disease process, in characterization of the viruses, in the development of the fields of immunology and diagnostic technologies, and in the establishment of regulations controlling the movement of production animals. Initial experiences confirmed that eradication of some infectious diseases from defined areas could be achieved with a test and slaughter program, even in the absence of an effective vaccine. For example, the recent global eradication of rinderpest was achieved through slaughter of infected animals, restriction of animal movement from enzootic areas to zones free of the infection, and vaccination of animals in enzootic regions. In this type of control program, individual animals could be sacrificed for the good of the production unit. With the increase in the importance of companion animals in today’s society, control programs based on depopulation of infected animals cannot be utilized simply because the individual animal is the important unit, as in human medicine.

 

Thus, in regular veterinary practice, canine parvovirus infections cannot be controlled by killing the affected animals and restricting the movement of dogs, and instead, effective vaccines must continue to be developed and utilized for prophylactic immunization. Diagnostic tests must be deployed that can rapidly detect infectious agents in a time frame such that the test results can direct treatment. As we become more aware of the interaction between domestic animals and wildlife, we also must face the reality that there are viruses transmitted by insect vectors that do not respect national boundaries and for which the range may be expanding because of climatic changes. Enhanced surveillance programs, novel and improved control strategies, and antiviral drugs will need to be developed continually in the future, particularly for those diseases for which vaccination is not yet possible or is not cost-effective. Viruses have traditionally been viewed in a rather negative context—disease-producing agents that must be controlled or eliminated. However, viruses have some beneficial properties that can be exploited for useful purposes. Specifically, some viruses (eg, baculovirus) have been engineered to express useful nonviral proteins or to express viral proteins for immunization purposes (eg, poxvirus and adenovirus vectored vaccines). Lentiviruses have been modified for the purpose of inserting genes of interest into cells for research purposes and for use in gene therapy, as have adeno-associated viruses (which are actually parvoviruses). Bacteriophages are being considered in the context of controlling certain bacterial infections, and viruses have the hypothetical potential to be vectors that selectively target tumor cells for controlling cancers. In the broader context of the Earth’s ecosystems, viruses are now viewed in a more positive sense, in that they may be a component of population control and perhaps a force in the evolution of species. Although restricting the population of agriculturally important animals is viewed as a negative from the human perspective, the ecosystem might benefit from the reduction of one species if its success is at the expense of others. An insect infestation that is curtailed by baculoviruses is considered to be beneficial, but the loss of poultry to influenza virus infection is viewed unfavorably even though the two events may be ecologically equivalent. We are now fully comfortable with the concept of beneficial bacteria in the ecosystem of the human body. Do we need to start to consider that viruses that have evolved with the species may also have beneficial properties?

 

CHARACTERISTICS OF VIRUSES

Following the initial operational definition of a virus as a filterable agent, attempts were made to identify properties of viruses that made them distinct from other microorganisms. Even from the earliest times it was evident that the filterable agents could not be cultivated on artificial media, and this particular characteristic has withstood the test of time, in that all viruses are obligate intracellular parasites. However, all obligate intracellular parasites are not viruses (Table 1.2).

 

 

Members of certain bacterial genera also are unable to replicate outside a host cell (eg, Ehrlichia, Anaplasma, Legionella, and Rickettsia). These “degenerate” bacteria lack key metabolic pathways, the products of which must be provided by the host cell. Viruses, by contrast, lack all metabolic capabilities necessary to reproduce, including energy production and the processes necessary for protein synthesis.

 

Viruses do not possess standard cellular organelles, such as mitochondria, chloroplasts, Golgi, and endoplasmic reticulum with associated ribosomes. However, cyanophages represent an exception as they encode proteins involved in photosynthesis that increase viral fitness by supplementing the host cell systems. Similarly, certain bacteriophages have genomes that encode enzymes involved in the nucleotide biosynthetic pathway. Outside the living cell, viruses are inert particles whereas, inside the cell, the virus utilizes the host cell processes to produce its proteins and nucleic acid to replicate itself. As will be noted later, the proteincoding capacity of viruses ranges from just a few proteins to several hundred. This range of complexity mirrors the diverse effects viral infections have on host cell metabolism, but the outcome of an infection is the same—the production of more progeny viruses.

 

A second inviolate property of viruses is that they do not reproduce by binary fission, a method of asexual reproduction in which a preexisting cell splits into two identical daughter cells; in the absence of limiting substrate, the population of cells will double with each replication cycle, and at all points in the replication cycle there exists a structure that is identifiable as an intact cell. For viruses, the process of reproduction resembles an assembly line in which various parts of the virus come together from different parts of the host cell to form new virus particles. Shortly after the virus attaches to a host cell, it enters the cell and the intact virus particle ceases to exist. The viral genome then directs the production of new viral macromolecules, which results ultimately in the assembly and appearance of new progeny virus particles. The period of time between the penetration of the virus particle into the host cell and the production of the first new virus particle is designated as the eclipse period, which varies depending on virus family. Disrupting cells during the eclipse period will interrupt the release of significant numbers of infectious virus particles. Uninterrupted, a single infectious particle can replicate within a single susceptible cell to produce thousands of progeny virus particles. As more sensitive analytical techniques became available and more viruses were identified, some of the criteria that defined a virus became less absolute. In general, viruses contain only one type of nucleic acid that carries the information for replicating the virus. However, is it now clear that some viruses contain nucleic acid molecules other than their genomic DNA or RNA. For retroviruses, cellular transfer (t)RNAs are essential for the reverse transcriptase reaction, and studies have shown that some 50
100 tRNA molecules are present in each mature virion. Similarly in herpesviruses, host cell and viral transcripts localize to the tegument region of the mature virion. Early studies defined viruses by their tiny size; however, “giant” viruses now have been identified that are physically larger than some mycoplasma, rickettsia, and chlamydia. The mimiviruses and pandoraviruses that infect amoeba are remarkable exceptions to existing rules: the mimivirus virion is approximately 0.75 μm (750 nm) in diameter, with a DNA genome of 1.2 megabases (nucleotides). The pandoraviruses are even larger (up to 1 μm) with a genome up to 2.5 megabases. Because of their virion size, these large viruses would be retained by standard 300-nm filters traditionally used for separating bacteria from viruses. The genomes of these giant viruses can include more than 1000 genes, including those encoding proteins potentially involved in protein translation, DNA repair, cell motility, and membrane biogenesis. For example, mimiviruses encode aminoacyltRNA synthetase that likely affords some independence from host cell pathways for replication of the viral genome. The discovery of these large viruses has revived the debate as to the origin of viruses. Furthermore, sequence data link mimiviruses to the nucleocytoplasmic large DNA viruses, specifically viruses in the families Poxviridae and Iridoviridae.

 

Chemical Composition of the Virion

The chemical composition of virus particles varies markedly between those of individual virus families. For the simplest of viruses such as parvoviruses (family Parvoviridae), the virion is composed of viral structural proteins and DNA, whereas in the case of picornaviruses (family Picornaviridae) it comprises viral proteins and RNA.

 

The situation becomes more complex with the enveloped viruses such as members of the Herpesviridae and Paramyxoviridae families. These types of viruses mature by budding through different host-cell membranes that are modified by the insertion of viral proteins. For the most part, host-cell proteins are not a significant component of viruses, but minor amounts of cellular proteins can be present in viral membranes and in the interior of the virus particle. Host-cell RNA such as ribosomal RNA can be found in virions, but there is no evidence for a functional role in virus replication.

 

For enveloped viruses, glycoproteins are the major type of protein present on the exterior of the membrane. The existence/presence of a lipid envelope provides an operational method with which to separate viruses into two distinct classes—those that are inactivated by organic solvents (enveloped) and those that are resistant (nonenveloped).

 

Viral Nucleic Acids in the Virion

Viruses exhibit remarkable variety with respect to genome composition and in strategies for the expression of their genes and for the replication of their genome. If one considers the simplicity of RNA plant viroids (247
401 nucleotides) at one extreme and the pandoraviruses (2.47 megabases) at the other, one might conclude that viruses have perhaps exploited all possible means of nucleic acid replication for an entity at the subcellular level. The type and structural characteristics of the viral genomic nucleic acids are used to classify viruses. As viruses contain only one nucleic acid type with respect to transmitting genetic information, the virus world can simply be divided into RNA viruses and DNA viruses (Fig. 1.1).

 

 

FIGURE 1.1 Diagrammatic representation of the spectrum of morphological types represented by animal viruses. From King, A.M., Adams, M.J., Carstens, E.B., Lefkowitz, E.J., (Eds.), 2012. Virus Taxonomy: Ninth Report of the International Committee on Taxonomy of Viruses. Elsevier Academic Press, San Diego, CA, p. 20. Copyright r Elsevier (2012), with permission.

 

For RNA viruses, one major distinction is whether the virion RNA is of positive sense or polarity, directly capable of translation to protein, or of negative sense or polarity, which requires transcription of the genome to generate mRNA equivalents. Within the negative-strand group, there are single-strand whole-genome viruses (eg, Paramyxoviridae) and segmented genome viruses (eg, Orthomyxoviridae—six, seven, or eight segments; Bunyaviridae—three segments; Arenaviridae—two segments). The Retroviridae are considered diploid, in that the virion contains two whole-genomic positive-sense RNAs. Some RNA viruses possess genomes comprised of double-stranded RNA. The Birnaviridae have two segments and the Reoviridae have 10, 11, or 12 segments, depending on the genus of virus. The size of animal RNA viral genomes ranges from less than 2 kilobases (kb) (Deltavirus) to more than 30 kb for the largest RNA viruses (Coronaviridae).

 

For the animal DNA viruses, the overall structure of the genomes is less complex, with either a single molecule of single-stranded (ss)DNA or a single molecule of double-stranded (ds)DNA. For the dsDNA viruses, the complexity ranges from the relatively simple circular super-coiled genome of the Polyomaviridae and Papillomaviridae (5
8 kbp) to the linear Herpesviridae (125
235 kbp) with variable sequence rearrangements. The ssDNA viral genomes are either linear (Parvoviridae) or circular (Circoviridae and Anelloviridae), with sizes ranging from 2.8 to 5 kbp. In general, the size of the viral genome influences the protein-coding capacity of the virus, but there is not a simple calculation that reliably estimates this relationship. Parts of the viral genome are typically regulatory elements necessary for the translation of viral proteins, replication of the genome, and transcription of viral genes (promoters, termination signals, polyadenylation sites, RNA splice sites, etc.). For specific examples and more detailed discussion the reader is referred to Chapter 2, Virus Replication.

 

Viral Proteins in the Virion

The genomes of animal viruses encode from as few as one protein to more than 100. Proteins that are present in virions (mature virus particles) are referred to as structural proteins, whereas proteins that are produced during the infection but are not incorporated into newly assembled virus particles are referred to as nonstructural proteins. Nonstructural proteins play essential roles in the virus replication process, such as regulating gene expression, replication of the genome, proteolytic processing of viral precursor proteins, facilitating the assembly of virus particles, or modification of the host innate response to infection.

 

There is some ambiguity for enzymes that are essential for the initial stages of virus replication, such as the RNA polymerases for the negative-strand RNA viruses (Paramyxoviridae, Rhabdoviridae, etc.). As the first step in the replication cycle, once the nucleocapsid enters the cytoplasm the viral genome is transcribed, requiring that the polymerase is part of the mature virion. Whether the polymerase has a true structural role in the mature particle in addition to its transcription activity is less certain. Numerous other viral proteins that occur within the virions of complex viruses (Poxviridae, Herpesviridae, Asfarviridae) also appear to have no apparent structural role. Virion proteins fall into two general classes: modified proteins and unmodified proteins.

 

The capsids of the nonenveloped viruses are composed of proteins with few modifications, as their direct amino acid interactions are essential for the assembly of the protein shells. Proteolytic cleavage of precursor proteins in the nascent capsid is not uncommon in the final steps of assembly of the mature capsid proteins. Glycoproteins are predominantly found in those viruses that contain a viral membrane. These structural proteins can be either a type I integral membrane protein (amino terminus exterior) (eg, hemagglutinin (HA) of influenza virus) or type II (carboxyl terminus exterior) (eg, neuraminidase of influenza virus). Glycosylation patterns may differ even amongst viruses that mature in the same types of cells, because N- and O-linked glycosylation sites on the virion proteins vary among the virus families.

 

The glycoproteins involved in virion assembly have a cytoplasmic tail that communicates with viral proteins on the inner surface of the membrane to initiate the maturation process for production of the infectious virus particle. Structural proteins in the infectious virus particle have a number of key functions: (1) to protect the genomic nucleic acid and associated enzymes from inactivation; (2) to provide receptor-binding sites for initiation of infection; and (3) to initiate or facilitate the penetration of the viral genome into the correct compartment of the cell for replication. The virion—that is, the complete virus particle—of a simple virus consists of a single molecule of nucleic acid (DNA or RNA) surrounded by a morphologically distinct capsid composed of viral protein subunits (virus-encoded polypeptides).

 

The protein subunits can self-assemble into multimer units (structural units), which may contain one or several polypeptide chains. Structures without the nucleic acid can be detected and are referred to as empty capsids. The meaning of the term nucleocapsid can be somewhat ambiguous. In a strict sense, a capsid with its nucleic acid is a nucleocapsid, but for simple viruses such as poliovirus, this structure is also the virion. For flaviviruses, the nucleocapsid (capsid 1 RNA) is enclosed in a lipid envelope and the nucleocapsid does not represent the complete virion. For paramyxoviruses, the nucleocapsid refers to a structure composed of a single strand of RNA complexed to a viral protein that assembles in the form of an α helix. The nucleocapsid assembles into a complete virion by obtaining a lipid envelope from host cell membranes modified by the insertion of viral proteins.

 

Viral Membrane Lipids

For viruses that mature by budding through a cellular membrane, a major constituent of the virion is a phospholipid bilayer that forms the structural basis of the viral envelope. The maturation site for viruses can be the plasma membrane, nuclear membrane, Golgi, or the endoplasmic reticulum. For those viruses budding from the plasma membrane, cholesterol is a constituent of the viral membrane, whereas the envelopes of those viruses that bud from internal membranes lack cholesterol. The budding process is not random, in that specific viral glycoprotein sequences direct developing particles to the proper location within the inner membrane surface. In polarized cells—cells with tight junctions, giving the cell a defined apical and basal surface—virus budding will be targeted to one surface over the other. For example, in Madin
Darby canine kidney cells, influenza virus will bud on the apical surface, whereas vesicular stomatitis virus buds from the basal surface (see Fig. 2.13). The transmembrane domain of viral glycoproteins targets specific regions of the cellular membrane for budding. For influenza virus, budding is associated with “lipid rafts,” which are microdomains of the plasma membrane rich in sphingolipids and cholesterol.

 

VIRAL MORPHOLOGY

Early attempts to characterize viruses were hampered by the lack of appropriate technologies. A major advance in determining virus morphology was the development of negative-stain electron microscopy in 1958. In this procedure, electron-dense stains were used to coat virus particles and produce a negative image of the virus with enhanced resolution (Fig. 1.2). Fig. 1.1 depicts the spectrum of morphological types represented by animal viruses. Given the remarkable variation in the size of virions of different viruses, from as much as 1000 nm to as little as 20 nm, it is not surprising that there were inconsistencies noted in historic filtration studies. Advances in determining virus morphology at the atomic level came from studies initially using X-ray crystallography and then combining this technique with other structural techniques such as electron cryomicroscopy (cryo-EM). In this process, samples are snap frozen and examined at temperatures of liquid nitrogen or liquid helium (Fig. 1.3). Cryo-EM offered the advantage that the samples are not damaged or distorted in the process of analyzing the structure, as occurs with negativestain electron microscopy and X-ray crystallography. However, the individual images generated by this process are of lower resolution than those obtained with crystallography. Critical to these analyses were the developments in computer hardware and software that were able to capture, analyze, and construct the threedimensional images from literally thousands of determinations. This “averaging” process can only work if the virus particles are uniformly the same size and shape. For many viruses, this uniformity is met by having the symmetry of a type of polyhedron known as an icosahedron. For intact virus particles showing icosahedral symmetry, the physical location of the individual peptides could be identified and the areas of the folded peptides that are on the surface of the virion were mapped. These areas could be linked to the specific epitopes recognized by monoclonal antibodies. In other studies, the binding site on the virion for the cellular receptor was mapped, which opened the possibility of developing antiviral drugs targeting these defined areas. X-ray crystallography can also be used to analyze subunits of a virus, such as was done for the HA protein of influenza virus (Fig. 1.4). The impact of mutations in the HA peptide as they related to changes in the binding of antibodies or host-cell receptors could be determined with these advanced technologies. Viruses come in a variety of shapes and sizes that depend on the shape, size, and number of their protein subunits and the nature of the interfaces between these subunits (Fig. 1.1). However, only two kinds of symmetry have been recognized in virus particles: icosahedral and helical. The symmetry found in isometric viruses is invariably that of an icosahedron; virions with icosahedron symmetry have 12 vertices (corners), 30 edges, and 20 faces, with each face an equilateral triangle. Icosahedra have two-, three-, and fivefold rotational symmetry, with the axes passing through their edges, faces, and vertices, respectively (Fig. 1.5). The icosahedron is the optimum solution to the problem of constructing, from repeating subunits, a strong structure enclosing a maximum volume. Parvoviruses represent one of the simplest capsid designs, being composed of 60 copies of the same protein subunit—three subunits per face of the icosahedron. The protein is folded into a structure referred to as a “jelly-roll β-barrel” that forms a block-like profile with an arm-like extension that provides the contact point with other subunits for stabilizing the protein
protein interactions. In the simplest arrangement, the size of the protein subunit determines the volume of the capsid. With a single capsid protein of 60 copies, only a small genome can be accommodated within the capsid (canine parvovirus 5 5.3 kb ssDNA). The explanations for the ways viruses maintain the icosahedron symmetry with repeating structural units is beyond the scope of this text. The nucleocapsid of several RNA viruses selfassembles as a cylindrical structure in which the protein structural units are arranged as a helix, hence the term helical symmetry. It is the shape and repeated occurrence of identical protein
protein interfaces of the structural units that lead to the symmetrical assembly of the helix. In helically symmetrical nucleocapsids, the genomic RNA forms a spiral within the core of the nucleocapsid. The RNA is the organizing element that brings the structural units into correct alignment. Many of the plant viruses with helical nucleocapsids are rod-shaped, flexible, or rigid without an envelope. However, with animal viruses, the helical nucleocapsid is wound into a secondary coil and enclosed within a lipoprotein envelope (Rhabdoviridae; Fig. 1.6). There are, inevitably, viruses that do not conform to the simple rules of morphology. For example, members of the Poxviridae have “complex” symmetry (see Chapter 7: Poxviridae). Similarly, there are highly pleomorphic viruses in which each virion has its own unique shape (eg, members of the Filoviridae, see Chapter 19: Filoviridae).

 

 

 

VIRAL TAXONOMY

With the earliest recognition that infectious agents were associated with a given spectrum of clinical outcomes, it was natural for an agent to take on the name of the disease with which it was associated or the geographic location where it was found, as there was no other basis for assigning a name. Thus the agent that caused foot-andmouth disease in cattle becomes “foot-and-mouth disease virus,” or an agent that caused a febrile disease in the Rift Valley of Africa became “Rift Valley fever virus.” It is not difficult at this time in history to see why this ad hoc method of naming infectious agents could lead to confusion and regulatory chaos, as different names may be given to the same virus. For example, hog cholera virus existed in North America whereas, in the rest of the world it was referred to as classical swine fever virus, not to be confused with African swine fever virus. Within the same animal, one had infectious bovine rhinotracheitis (IBR) virus and infectious bovine pustular vulvovaginitis (IBPV) virus—both disease entities being caused by bovine herpesvirus 1. Even today, export certification documents may ask for tests to certify animals free of IBR virus and IBPV virus. This disease-linked nomenclature could not be changed until such time as the tools became available to define the physical and chemical nature of viruses. With negative-stain electron microscopy as a readily available technology, the size and shape of viruses became a characteristic for defining them. This, along with the ability to define the type of nucleic acid in the virus particle, provided the beginnings of a more rational system of classifying and naming new viruses. Even with a defined shape and a type of nucleic acid, there were still ambiguities in the classification systems that were being developed. Viruses that were transmitted by insect vectors were loosely defined as “arboviruses”— arthropod-borne viruses. However, there were viruses that “looked like” arboviruses (togaviruses—viruses with a symmetrical lipid membrane) and had the same nucleic acid, but did not have an insect vector. These became “nonarthropod-borne” togaviruses. These ambiguities were increasingly resolved with access to the nucleotide sequences of these agents. Thus, for example, the “nonarbo” togaviruses became members of the genera Rubivirus, Pestivirus, and family Arteriviridae. Whereas viruses initially were classified according to the diseases they caused, shared physical and chemical properties, and serologic cross-reactivity, the advent of nucleic acid sequencing technologies developed in the molecular era allowed genetic comparisons of different viruses to facilitate taxonomic classifications. In general, genetic relationships parallel those previously established by the older criteria. Virus sequencing also allows for phylogenetic comparisons to determine the evolutionary development and history of viral species. This is a powerful tool for defining viral ancestries. However, a major limitation to sequence-based classification is that inferences are compromised by the variable nature of viruses, especially for highly divergent RNA viruses. Despite this, phylogenies that analyze the most conserved motifs of viral RNA dependent RNA polymerase sequences have been used to generate higher-order classifications that define viral “supergroups” and establish family-level distinctions. For example, phylogenetic analyses of viruses including retroviruses and hepadnaviruses with reverse transcriptase activity have been more informative than polymerase sequences due to a higher degree of sequence conservation of reverse transcriptase genes. New methods presently being developed to circumvent inferences from sequences will include comparing genome organization (eg, gene content and order) as well as protein secondary structure. The International Committee on Taxonomy of Viruses (ICTV) was established in 1966 to establish, refine, and maintain a universal system of virus taxonomy. Given the uncertain origins of viruses, establishing the initial framework for this classification system was not without controversy. Subcommittees and study groups meet periodically to assess new data submitted from the research community to refine the classification system and to place new viruses in their most logical position in the taxonomy scheme. It was not until the Seventh Report of the ICTV (2000) that the concept of virus species as the lowest group in the viral taxa was accepted. The advent of nucleotide sequence determination had a dramatic effect on all biological classification systems, and it has in many respects confirmed the major elements of the classification system. As the process of classification and defining nomenclature is an ongoing one because of the discovery of new viruses and the generation of sequence data on historic virus isolates, it is impossible for a textbook to ever be truly “current.” This textbook will use the information presented in the Ninth Report of the ICTV published in 2011, as updated by the ICTV online resource (http://www.ictvonline.org/ index.asp). The hierarchy of recognized viral taxa is: Order; Family; (Subfamily); Genus; Species. For example, human respiratory syncytial virus A2 would be found in this system as: Mononegavirales (order); Paramyxoviridae (family); Pneumovirinae (subfamily); Pneumovirus (genus); Human respiratory syncytial virus (species). The 2011 ICTV report lists 2284 species of virus and viroid distributed amongst 349 genera, 19 subfamilies, 87 families, and 6 orders. To be a member of the taxa higher than species, a virus must have all properties defining the classification. In contrast, species are considered a polythetic class, in which members have several properties in common but all do not have to share a single defining property. For each genus, there has been designated a type species, which is a species that creates a link between the genus and the species. This designation is usually conferred on the species that necessitated the creation of the genus. The published virology literature contains obvious inconsistencies with regard to whether the name of a specific virus is capitalized and/or written in italics: Bovine viral diarrhea virus versus bovine viral diarrhea virus, for example. In all cases dealing with taxonomy, the order, family, subfamily and genus names should be written in italics and capitalized. In discussing a virus in the context of taxonomy at the species level, the name is written in italics and the first word is capitalized: for example, Canine distemper virus is a species in the genus Morbillivirus. However, when a virus is written about in terms of tangible properties such as its ability to cause disease, growth in certain cell lines, or its physical characteristics, the name is neither written in italics nor capitalized unless the name contains a proper noun; for example, one can grow canine distemper virus or West Nile virus in monkey cells. There are instances when the abstract (taxonomy) and the concrete aspects of a virus are not clear in the context of the sentence. In this textbook we will attempt to use the ICTV conventions when clearly appropriate, but as this text deals mainly with the tangible aspects of viruses, most virus names will not be in italics. A basic question that has yet to be addressed is why we should bother with taxonomy at all. For some there seems to be a human need to place things into an ordered system. In characterizing an entity and defining a nomenclature, a basic understanding of the subject under study may be achieved. In a larger context, taxonomy provides a tool for comparing one virus with another or one virus family with another. It also enables one to assign biological properties to a new virus that is provisionally linked to a given family. For instance, if one has an electron micrographic image of a new virus that supports its identity as a coronavirus, then the discoverer can assume they have identified a single-stranded, positive-sense, nonsegmented RNA virus. Further, one can extrapolate that coronaviruses are mainly associated with enteric disease, but can also cause respiratory disease in “atypical” hosts after “species jumping.” As a group, coronaviruses are difficult to culture in vitro, and may require the presence of a protease to enhance growth in tissue culture. Conserved sequences—perhaps in the nucleocapsid—might provide a target for the development of a PCR test. Thus, identification of the morphology of an unknown virus can be useful, as the general properties of specific virus families can assist in the interpretation of individual clinical cases. For example, confirming that an alphaherpesvirus was isolated from a particular case, or its presence identified by deep sequence analysis of clinical material, confers some basic knowledge about the virus without having explicitly to define the properties of the specific virus species responsible. However, current taxonomy of viruses is not without confusion. There is substantial variation in how viruses are classified currently within individual families; for example, viruses in the family Flaviviridae (genus Flavivirus) are still grouped according to their serological relationships, whereas viruses in the family Picornaviridae are increasingly subdivided into genera based on their genome sequences and organization. Furthermore, the designation of a “virus species” can include a variety of other “virus isolates” such that, despite their very different biological properties and host range (species tropism), feline panleukopenia virus and canine parvovirus 2 are both representatives of Carnivore protoparvovirus 1 (see Chapter 12: Parvoviridae). More detailed properties of the virus families that include significant pathogens of veterinary relevance will be found in specific chapters in Part II of this text.

 

Phylogenetic Comparison of Virus Sequences

Prior to the advent of molecular biology, viruses were classified according to their serological relationships. Sequencing technologies developed in the molecular era have allowed for genetic comparison of viruses, which generally match relationships that were previously serologically defined. Phylogenies are tree-like pictorial descriptions of the evolutionary history of a particular virus species or family, where each branch tip represents a specific virus sequence; these “trees” are usually generated based on sequence comparison of the most conserved region of the viral genome (for representative examples, see Figs. 17.1, 18.1, and 21.2). Although inferences based on viral sequence data are compromised by the variable nature of viruses, especially for highly divergent RNA viruses, phylogenetics has proven useful for generating higher-order classifications to define “supergroups” of viruses. Beyond its importance to virus taxonomy, the advent of molecular virology ushered in a new era for the study of virus evolution via phylogenetic comparisons. This approach was used, amongst many examples, to determine that HIV originated from SIV that infect nonhuman primates. A phylodynamics approach can also be used to infer the origins, epidemiology, and dynamics of viruses during epidemics. By comparing gene sequences of viruses and analysis of phylogenetic trees, valuable information can be derived regarding virus population growth and decline, the extent of population subdivision, and viral migration. These approaches prove especially useful for RNA viruses that change rapidly, allowing the resolution of phylogenetic relationships between samples obtained only days apart for example. They are also used to document dispersal of specific viruses, such as influenza viruses amongst birds and humans. In most cases, phylogenies only show evolutionary order and not the length of time between two sequences, unless special molecular clock models are applied. For viruses that recombine or reassort, valid phylogenies also require analysis of multiple genomic regions. Phylogenetic analyses can take many forms, although they invariably result in the generation of a phylogenetic tree. Neighbor joining tree-building algorithms calculate the genetic distance, measured via a matrix, between each pair of viral sequences being compared; the resulting topology minimizes the distance between nearest neighbors. This approach is rapid and therefore favored for generating a tentative tree or for choosing the best tree among multiple options, but since sequence data is reduced to a distance matrix at the outset, if the matrix is incorrect a false tree may be produced. Maximum parsimony is a nonparametric statistical method for producing trees where branches are placed in the simplest way possible to support minimal evolutionary change. Parsimony algorithms are best used when viruses share high genetic conservation and when the number of sequences being analyzed is low since the method is time-intensive. The approach is not always guaranteed to produce a true tree with high probability, especially when evolution is rapid. A third approach, maximum likelihood, which uses a parametric statistical model to provide estimates for the parameters in the model and then determines the probability of observing the tree topology given the model, is even slower than parsimony, but is favored for confirming trees built using other algorithms, especially with small data sets since it is less affected by sampling error as compared to the other methods. Trees generated with all three algorithms rely on bootstrap analyses, typically reported at branch nodes, to provide statistical support for their topologies. Bootstrap values of 95 or higher are statistically robust and denote that if the tree was rebuilt 100 times using the same method, the same relative positions of the sequences at the node would occur 95 out of 100 times.